Classification of forearm and finger motions using electromyogram and arm-shape-changes

Area of Science:

• Robotics
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Robot arms for humanoid applications are common in medical, welfare, and education.
• Surface electromyogram (sEMG) signals are used for hand motion classification but struggle with detailed movements.
• Existing motion capture technologies like Kinect have limitations in setup and detection range.

Purpose of the Study:

• To develop an advanced method for classifying detailed human motions.
• To overcome the limitations of using sEMG or Kinect alone for motion recognition.
• To enhance the classification accuracy of intricate hand and wrist movements.

Main Methods:

• Proposed a novel method combining surface electromyogram (sEMG) signals with arm-shape changes.
• Defined arm-shape changes as forearm deformation due to muscle bulging during movement.
• Utilized a combined approach to capture both electrical and physical motion data.

Main Results:

• Achieved classification accuracies of 90% or more for wrist pronation and supination.
• Demonstrated superior performance in classifying detailed motions compared to sEMG alone.
• Successfully classified intricate movements previously difficult to discern.

Conclusions:

• The proposed combined method effectively classifies detailed hand and wrist motions.
• This approach significantly expands the range of classifiable hand motions for robotic applications.
• The findings contribute to more sophisticated human-robot interaction and control systems.